Ansys non-linear material simulation

[MMilis]

Hello all

Any tips hints directions to solving my problem are much appreciated!

I am trying to model a composite sandwich structure using a orthotropic core material that has a non-linear deformation curve. I am using TB,PLASTIC,,,,MISO and so far I am getting results, however I don't trust my results. The non-linear table I input gets surpassed by quite a bit. It is defined up to a strain of 7.32%, for which the stress is 1.57MPa and at which the stiffness slope is 0. I would have expected Ansys to stick to this defined material data, however it's computed stresses and strains for elements with these material properties are up to 70$ and 6MPa stress.

I don't quite know where to look for a solution to this. The fact that strains go outside of the defined region seems normal since the final stiffness is zero. However I feel they should still remain at their maximum defined stress. I will try to use TB,MISO however I am pretty sure this will not have a significant effect on the results. If this doesn't work, I think I will try to 'kill' the elements once they reach a certain stress/strain value, however I was hoping there was an easier way to obtain a realistic stress-strain distribution in my model.

Thanks a lot for your time!

I have defined my plastic material in the following way;
et,2,solid185
MP,EX,2,45
MP,EY,2,45
MP,EZ,2,45
MP,GXY,2,22
MP,GYZ,2,22
MP,GXZ,2,22
MP,PRXY,2,0.42
MP,PRYZ,2,0.42
MP,PRXZ,2,0.42
TB,PLASTIC,2,1,26,MISO
TBTEMP,0
TBPT,DEFI, 0.000000000000000 , 1.003048937000000
TBPT,DEFI, 0.00162033300000000 , 1.073777736000000
TBPT,DEFI, 0.00360120700000000 , 1.125693844000000
TBPT,DEFI, 0.00504243800000000 , 1.156856602000000
TBPT,DEFI, 0.00684421000000000 , 1.193219388000000
TBPT,DEFI, 0.00900699100000000 , 1.229600879000000
TBPT,DEFI, 0.01116977300000000 , 1.265982370000000
TBPT,DEFI, 0.01423461000000000 , 1.307591947000000
TBPT,DEFI, 0.01693937200000000 , 1.338820173000000
TBPT,DEFI, 0.01910355700000000 , 1.359657696000000
TBPT,DEFI, 0.02126774100000000 , 1.380495218000000
TBPT,DEFI, 0.02487549400000000 , 1.406588884000000
TBPT,DEFI, 0.02776169700000000 , 1.427463817000000
TBPT,DEFI, 0.03136991800000000 , 1.448376160000000
TBPT,DEFI, 0.03353503700000000 , 1.458851037000000
TBPT,DEFI, 0.03624167100000000 , 1.469353971000000
TBPT,DEFI, 0.03894830400000000 , 1.479856906000000
TBPT,DEFI, 0.04237648800000000 , 1.495578573000000
TBPT,DEFI, 0.04724870900000000 , 1.511375061000000
TBPT,DEFI, 0.05103836900000000 , 1.521934111000000
TBPT,DEFI, 0.05518903900000000 , 1.532511866000000
TBPT,DEFI, 0.06078327800000000 , 1.548345764000000
TBPT,DEFI, 0.06565596600000000 , 1.558960930000000
TBPT,DEFI, 0.07034861700000000 , 1.564385419000000
TBPT,DEFI, 0.07323622300000000 , 1.569716383000000
TBPT,DEFI, 0.0735 , 1.569716383000000

 

[L_K]

Not knowing a lot of composite modelling but what stress component is going well over the defined limit?
Also at least when using isotropic material, stresses can be higher than defined limit because of stress interpolation from gauss point to nodes.
This effect can be at least seen on surfaces. It should not however explain such high stresses.

 

[MMilis]

L_K, thanks for your quick response.

I am modelling a simple uni-axial test where the sandwich is loaded in out-of plane tension (z-direction). The stress in z-direction is going well over the defined limit of 1.5697.

This is difficult to explain but I'll try; (see also attached picture)
The core of the sandwich structure I am modelling has an incorporated grid formed by epoxy 'walls' that extends almost through the thickness. This occurs due to usage of grid scored core material of which the grid gets filled by epoxy resin during infusion production process.

Best regards
Maarten

 

[scheah]

Hi Maarten,

Yes, the stiffness is zero past the last point on your table. A way to keep the same gradient beyond that is to add another point on your table that keeps the same gradient at a much higher strain.


Best regards,
Sze Kwan (Jason) Cheah

 

[MMilis]

Hi sk_cheah

Thank you for your response. I already tried adding more points to the graph, and I'm not really worrying about strains becoming really large, however the corresponding stresses go way beyond the graph that I have defined as input. This is what worries me.

 

[L_K]

Could you post the picture of the setup again.